Collaborative Development of an Institution-Wide Controlled Vocabulary

What this was about

Caitlin Thorn, a data manager at the Museum für Naturkunde Berlin, described how an interdisciplinary working group built a shared controlled vocabulary for about 30 million objects whose sub-collections used different terms for the same thing (object type, preparation type, availability). Instead of a dedicated thesaurus system they chose a structured Excel spreadsheet: collection staff already know Excel, and the list can move to a dedicated system later. The work ran from January 2024 workshops through a Collection Cafe presentation and specialist review to implementation as Specify pick lists. Q&A covered hierarchies, whether the list will be shared, governance, and alignment with Darwin Core.

Why it matters. Many institutions migrating to a shared collection management system hit the same terminology problem. This talk gives a low-tech, people-centred recipe that others (Naturalis said it is about to try the same) can reuse.

In the room

• Collections used different terms for the same thing, or the same term for different things, e.g. for object type, preparation type and availability.
• A thesaurus was treated simply as a list of terms with definitions; the group's main effort went into writing disambiguating definitions (e.g. 'lost' means permanently gone, 'missing' means it may reappear).
• Commercial thesaurus systems were evaluated, but a structured Excel sheet (English, German, Latin where relevant, definitions in two languages) was chosen for accessibility.
• The internal weekly 'Collection Cafe' forum and institution-wide emails were used to build participation and credibility.
• The final vocabulary was aligned with Darwin Core lists and implemented as Specify pick lists, with a process for requesting changes.
• The group has no authority over collections: nearly every term a collection asked for was included, and only ambiguous terms were negotiated.

Notable moments